Added a minimum word length config option, "wordlen_min".

Also did some minor cleanup of the word add code in process_file.

3 files changed, 22 insertions, 23 deletions

diff --git a/config.c b/config.c
index 0ae6beb..2cb6aa6 100644
--- a/config.c
+++ b/config.c
@@ -35,6 +35,10 @@ int cfg_init() {
 		ircstats_config.monolog_min = 5;
 	}
 
+	if(!config_lookup_int(&config, "wordlen_min", &ircstats_config.wordlen_min)) {
+		ircstats_config.wordlen_min = 3;
+	}
+
 	config_setting_t *regexes_setting = config_lookup(&config, "regexes");
 	if(!config_setting_is_aggregate(regexes_setting)) {
 		fprintf(stderr, "Setting \"regexes\" must be an aggregate type.\n");
diff --git a/config.h b/config.h
index 3b42ee8..775d7eb 100644
--- a/config.h
+++ b/config.h
@@ -5,7 +5,7 @@ int cfg_init();
 void cfg_free();
 
 struct ircstats_config_t {
-	long int threads, monolog_min;
+	long int threads, monolog_min, wordlen_min;
 };
 
 extern struct ircstats_config_t ircstats_config;
diff --git a/parsing.c b/parsing.c
index 1172cd1..4ef3fb7 100644
--- a/parsing.c
+++ b/parsing.c
@@ -21,6 +21,19 @@ static pthread_mutex_t user_mutex, word_mutex, channel_mutex;
 static struct user_t *last_user = NULL;
 static int in_monolog = 0, monolog_len = 0;
 
+static inline void add_word(struct user_t *user, wchar_t *word, int len) {
+	pthread_mutex_lock(&user_mutex);
+	user->words++;
+	pthread_mutex_unlock(&user_mutex);
+	word[len] = '\0';
+	char mbword[TEXT_BUFFER_SIZE];
+	wcstombs(mbword, word, TEXT_BUFFER_SIZE);
+	pthread_mutex_lock(&word_mutex);
+	struct word_t *word_s = word_get(mbword);
+	word_s->count++;
+	pthread_mutex_unlock(&word_mutex);
+}
+
 static void process_file(FILE *f, struct channel_t *channel, struct regexset_t *rs) {
 	char line[LINE_BUFFER_SIZE];
 
@@ -79,17 +92,8 @@ static void process_file(FILE *f, struct channel_t *channel, struct regexset_t *
 			int len = 0;
 			for(wchar_t *pos = wtext; pos < end; pos++) {
 				if(iswblank(*pos)) {
-					if(len) {
-						pthread_mutex_lock(&user_mutex);
-						user->words++;
-						pthread_mutex_unlock(&user_mutex);
-						word[len] = '\0';
-						char mbword[TEXT_BUFFER_SIZE];
-						wcstombs(mbword, word, TEXT_BUFFER_SIZE);
-						pthread_mutex_lock(&word_mutex);
-						struct word_t *word_s = word_get(mbword);
-						word_s->count++;
-						pthread_mutex_unlock(&word_mutex);
+					if(len >= ircstats_config.wordlen_min) {
+						add_word(user, word, len);
 					}
 					len = 0;
 					*word = '\0';
@@ -100,17 +104,8 @@ static void process_file(FILE *f, struct channel_t *channel, struct regexset_t *
 					*word = '\0';
 				}
 			}
-			if(len) {
-				pthread_mutex_lock(&user_mutex);
-				user->words++;
-				pthread_mutex_unlock(&user_mutex);
-				word[len] = '\0';
-				char mbword[TEXT_BUFFER_SIZE];
-				wcstombs(mbword, word, TEXT_BUFFER_SIZE);
-				pthread_mutex_lock(&word_mutex);
-				struct word_t *word_s = word_get(mbword);
-				word_s->count++;
-				pthread_mutex_unlock(&word_mutex);
+			if(len >= ircstats_config.wordlen_min) {
+				add_word(user, word, len);
 			}
 			continue;
 		}